contToDisc: Continuous to Discrete Transformation

Description Usage Arguments Value Note Author(s) See Also Examples

View source: R/DiscSurvDataTransform.R

Description

Discretizes continuous time variable into a specified grid of censored data for discrete survival analysis. It is a data preprocessing step, before the data can be extendend in long format and further analysed with discrete survival models.

Usage

1
contToDisc(dataSet, timeColumn, intervalLimits, equi=FALSE)

Arguments

dataSet

Original data in short format. Must be of class "data.frame".

timeColumn

Name of the column with discrete survival times. Must be a scalar character value.

intervalLimits

Numeric vector of the right interval borders, e. g. if the intervals are [0, a_1), [a_1, a_2), [a_2, a_max), then intervalLimits = c(a_1, a_2, a_max)

equi

Specifies if argument *intervalLimits* should be interpreted as number of equidistant intervals. Logical only TRUE or FALSE is allowed.

Value

Gives the data set expanded with a first column "timeDisc". This column includes the discrete time intervals (ordered factor).

Note

In discrete survival analysis the survival times have to be categorized in time intervals. Therefore this function is required, if there are observed continuous survival times.

Author(s)

Thomas Welchowski [email protected]

See Also

dataLong, dataLongTimeDep, dataLongCompRisks

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
# Example copenhagen stroke study data
library(pec)
data(cost)
head(cost)

# Convert observed times to months
# Right borders of intervals [0, a_1), [a_1, a_2), ... , [a_{\max-1}, a_{\max})
IntBorders <- 1:ceiling(max(cost$time)/30)*30

# Select subsample
subCost <- cost [1:100, ]
CostMonths <- contToDisc (dataSet=subCost, timeColumn="time", intervalLimits=IntBorders)
head(CostMonths)

# Select subsample giving number of equidistant intervals
CostMonths <- contToDisc (dataSet=subCost, timeColumn="time", intervalLimits=10, equi=TRUE)
head(CostMonths)

discSurv documentation built on May 29, 2017, 6:47 p.m.